Research On Knowledge Flow Network Based On Paper Citation Network

In recent years,with the rapid development of the Internet and the popularization of computers,our society has entered an era of knowledge explosion.The formation and growth of knowledge is a dynamic process,embodied in the flow of knowledge transferred between social network nodes.Therefore,analyzing the relationship between knowledge flow and mining knowledge in massive texts has become a research hotspot in data mining.In this thesis,the citation network is selected as the path carrier of knowledge flow propagation,and two data sets such as Artificial Intelligence and the Lancet are used to collect data and construct citation network respectively.The reference networks are divided into different research directions based on GN algorithm,and the results are verified based on knowledge flow as a data pre-processing for the knowledge flow study.In this thesis,the knowledge flow in the citation network of the paper is investigated.With the help of keywords in citation sentences,this thesis proposes two indicators to measure the intensity of knowledge flow: text similarity and noun distribution rate,and constructs a knowledge flow network based on this.In the knowledge flow network,different nodes carry different knowledge.Through the flow of knowledge,this thesis proposes a node model that plays an important role in the knowledge transfer in the knowledge flow network and experimental results verify it.It finds that the knowledge flow of the Lancet data set is stronger than that of the Artificial Intelligence data set.This thesis proposes a knowledge flow traceability method to classify the references in a paper by sorting out the knowledge sources of the paper and expanding the citations.Based on the method proposed in this research,the knowledge flow in the paper citation network can be discovered,and the depth of knowledge can be mined,which can help readers to quickly track valuable references.